DOI | Trouver le DOI : https://doi.org/10.1109/ISBI56570.2024.10635405 |
---|
Auteur | Rechercher : Tranchon, Antonin; Rechercher : Kunz, Manuela1; Rechercher : Séoud, Lama |
---|
Affiliation | - Conseil national de recherches du Canada. Technologies numériques
|
---|
Bailleur de fonds | Rechercher : National Research Council of Canada |
---|
Format | Texte, Article |
---|
Conférence | 2024 IEEE International Symposium on Biomedical Imaging (ISBI), May 27-30, 2024, Athens, Greece |
---|
Sujet | image segmentation; MRI; vertebrae; spine; scoliosis; watershed |
---|
Résumé | Adolescent Idiopathic Scoliosis (AIS) is a complex 3D deformity of the spine that requires surgical treatment for the most severe cases. Magnetic resonance imaging (MRI) of the spine is used for surgical planning and provides a reliable 3D pre-operative model to which intra-operative images can be registered for surgery assistance. In this work, we propose a novel approach for segmenting whole vertebrae in MRI of patients with severe scoliosis. It consists in a Unet segmentation model trained on MRIs of AIS patients that is further refined through a marker-controlled watershed algorithm. The external markers are obtained by aligning a template vertebra onto the initial segmentation. Results on 50 MRI volumes of preoperative AIS patients demonstrate an overall Dice score of 84%, but most importantly, an improvement of 47% at the level of the vertebral arch attributed to the refinement step. While most previous works have focused on segmenting the vertebral body, this work improves the segmentation of the vertebral processes, which are more challenging due to their irregular shape. |
---|
Date de publication | 2024-08-22 |
---|
Maison d’édition | IEEE |
---|
Dans | |
---|
Langue | anglais |
---|
Publications évaluées par des pairs | Oui |
---|
Exporter la notice | Exporter en format RIS |
---|
Signaler une correction | Signaler une correction (s'ouvre dans un nouvel onglet) |
---|
Identificateur de l’enregistrement | c86ba2ca-0227-40ff-9ef2-8cdcbc639704 |
---|
Enregistrement créé | 2024-08-28 |
---|
Enregistrement modifié | 2024-09-03 |
---|